Real-Time Video Process Monitoring for Low-Resource Abnormality Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing process monitoring systems face challenges in accurately classifying production process steps and detecting abnormalities in real-time due to resource-intensive image processing and lack of classification of process steps, especially when the same operation is repeated.
Innovation Solution
A deep learning-based real-time process monitoring system that classifies objects into moving, status, and vector objects, extracts features from real-time video using AI image classification neural networks, and compares these features with pre-stored patterns to detect abnormalities and classify processes, enabling efficient resource usage and accurate monitoring.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If image processing is performed to monitor the overall process and classify process steps, then process monitoring capability is improved, but resource consumption increases significantly
Solution Approach 1:
The patent segments the process monitoring task by dividing process steps into templates representing typical operation sequences. Instead of processing all images uniformly, the system creates separate processing paths for different process steps, allowing efficient resource allocation to each segment based on its specific needs.
Solution Approach 2:
The system performs preliminary actions by pre-defining process step templates that represent typical operation sequences before actual monitoring begins. These templates are created in advance and stored for rapid comparison during runtime, eliminating the need for complex real-time classification and reducing computational resource consumption.
2Measurement precision
If process steps are classified to determine which step an image corresponds to, then monitoring accuracy is improved, but processing complexity increases
Solution Approach 1:
The patent creates simplified copies of process step characteristics in the form of templates. These templates capture the essential features of each process step without requiring complex real-time analysis, allowing accurate classification through straightforward comparison between incoming images and pre-stored template patterns.
3Adaptability or versatility
If real-time video analysis is performed without process controller communication, then system independence is improved, but detection accuracy may deteriorate
Solution Approach 1:
The system performs self-service by independently analyzing video frames to detect process abnormalities without requiring communication with external process controllers. It uses its own embedded knowledge base of process step templates to autonomously classify images and identify deviations, maintaining both independence and accuracy.
Data Source
AI summary
The present invention relates to a deep-learning-based real-time process monitoring system and method, which register and learn an object to be recognized in a process, detect features of the object from a real-time video through classification of the object into a moving object, a status object, and a vector object based on a trained model, and monitor a progress state of the process through classification of an actually progressing process according to the features, thereby enabling easy detection of an abnormality of the process or the object while achieving improvement in performance of process monitoring and abnormality detection by processing the real-time video with small resources.